on whether the starting point is substance vs. form. In the first approach one asks how different
systems exploit a particular phonetic property. In the examples cited, it was seen that obstruent
voicing can be distinctive, demarcative or allophonic. The possibilities can be more extensive,
as in the case of nasality. As summarized below, there are at least five possibilities for how
nasality may be underlying contrastive in a language (cf. Cohn 1993, Clements & Osu 2003):
(1) a. on consonants only:
/m, n,
ŋ
/
e.g. Korean
b. on vowels and consonants:
/
᷉
i
,
᷉
u, ã, m, n,
ŋ
/
e.g. Bambara
c. on vowels only:
/
᷉
i
,
᷉
u, ã/
e.g. Ikwere
d. on whole morphemes:
/CVC/
N
e.g. Desano
e. absent entirely:
-----
e.g. Doutai
In addition to the above distinctions, languages may vary in whether they contain voiceless
nasals, prenasalized or nasally released consonants, as well in whether the contrasts are found
on all nasalizable consonants (e.g. including liquids and glides) and on all vowels. Similar
substance-directed typology can be done with virtually any phonetic feature or property, e.g.
voicing, aspiration, rounding, and so forth (cf. (2)). Still being substance-directed, a typologist
will likely be interested in how one vs. another of these properties is distributed in the
languages of the world, whether by genetic affiliation or by geography.
The second approach to phonological typology is form-directed: In this case the analyst
explores the logical properties of a specific model. The above examples from Trubetzkoy fall
into this category, as he was interested in the logical differences in the nature of the contrasts
that his model of phonology recognized. It mattered less that /l/ and /r/ differed in laterality or
rhoticity than the fact that they constitute an isolated bilateral contrast in any language which
has only these two liquids. This second, form-directed approach finds reincarnation in virtually
every model, if not every proposal in phonological theory. Since early generative phonology
proposed ordered rules (Chomsky & Halle 1968), it was only natural that a form-directed
typology should develop how these rules apply to forms and, in so doing, how they affect each
other: A phonological rule could apply to a form left-to-right, right-to-left, simultaneously, and
cyclically or non-cyclically. Earlier applying rules could be in feeding, bleeding,
counterfeeding, and counterbleeding relationships, creating and/or taking away inputs to which
later rules could apply (Kiparsky 1968, Kenstowicz & Kisseberth 1977). More recently, within
optimality theory (Prince & Smolensky 1993), all of the possible rankings can be exhaustively
computed in a “factorial typology” (cf. Gordon 2007). In short, most any formal property can
be “typologized” in terms of its logical parameters.
2. Phonology vs. typology
In both of the above approaches to phonological typology there has been a deep commitment to
the idea that phonetics and phonology are distinct from each other. As Buckley (2000: 2) puts
it, “... becoming divorced from the phonetics is the very essence of phonology.” The key goal
of phonology has been to determine what is a possible phonological system. This has meant
both determining the universal properties of sound patterns in languages as well as what’s
going on in the heads of speakers with respect to these sound patterns. While these goals are
directed towards the quest for universals, the traditional approach has been to seek universals
through the study of language particulars, which can be quite diverse. Determining how
Do'stlaringiz bilan baham: